Electrotrophy as potential primary metabolism for colonization of conductive surfaces in deep-sea hydrothermal chimneys

HAL Id: hal-03022667

https://hal-amu.archives-ouvertes.fr/hal-03022667

Preprint submitted on 24 Nov 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

hydrothermal chimneys

Guillaume Pillot, Sylvain Davidson, Laetitia Shintu, Oulfat Amin, Anne Godfroy, Yannick Combet-Blanc, Patricia Bonin, Pierre-Pol Liebgott

To cite this version:

Guillaume Pillot, Sylvain Davidson, Laetitia Shintu, Oulfat Amin, Anne Godfroy, et al.. Electrotrophy as potential primary metabolism for colonization of conductive surfaces in deep-sea hydrothermal chimneys. 2020. �hal-03022667�

1

Electrotrophy as potential primary metabolism for colonization of

1

conductive surfaces in deep-sea hydrothermal chimneys.

2

Guillaume Pillot¹, Sylvain Davidson¹, Laetitia Shintu³, Oulfat Amin Ali^X, Anne Godfroy², Yannick Combet-Blanc¹, Patricia

3

Bonin¹, Pierre-Pol Liebgott^1*

4

1 Aix Marseille Univ., Université de Toulon, IRD, CNRS, MIO UM 110, 13288, Marseille, France

5

2 IFREMER, CNRS, Université de Bretagne Occidentale, Laboratoire de Microbiologie des Environnements Extrêmes –

6

UMR6197, Ifremer, Centre de Brest CS10070, Plouzané, France.

7

3Aix Marseille Univ, CNRS, Centrale Marseille, iSm2, Marseille, France

8

*To whom correspondence may be addressed. Email: pierre-pol.liebgott@mio.osupytheas.fr; Mediterranean Institute

9

of Oceanography, Campus de Luminy, Bâtiment OCEANOMED, 13288 Marseille Cedex 09.

10

https://orcid.org/0000-0002-2559-1738

11

Key-words

12

Electrotrophy, hyperthermophiles, microbial electrochemical system, deep-sea hydrothermal 13

vent, electrosynthesis 14

Summary

15

Deep-sea hydrothermal vents are extreme and complex ecosystems based on a trophic chain. We 16

are still unsure of the first colonizers of these environments and their metabolism, but they are 17

thought to be (hyper)thermophilic autotrophs. Here we investigate whether the electric potential 18

observed across hydrothermal chimneys could serve as an energy source for these first colonizers.

19

Experiments were performed in a two-chamber microbial electrochemical system inoculated with 20

deep-sea hydrothermal chimney samples, with a cathode as sole electron donor, CO2 as sole 21

carbon source, and three different electron acceptors (nitrate, sulfate, and oxygen). After a few 22

days of culture, all three experiments showed growth of an electrotrophic biofilm consuming 23

directly or indirectly the electrons and producing organic compounds including acetate, glycerol, 24

2 and pyruvate. The only autotrophs retrieved were members of Archaeoglobales, in all three 25

experiments. Various heterotrophic phyla also grew through trophic interactions, with 26

Thermococcales in all three experiments and other bacterial groups specific to each electron 27

acceptor. This electrotrophic metabolism as energy source to drive the first microbial colonization 28

of conductive hydrothermal chimneys was discussed.

29

Introduction

30

Deep-Sea hydrothermal vents, are geochemical structures housing an extreme ecosystem rich in 31

micro- and macro-organisms. Since their discovery in 1977 (Corliss and Ballard, 1977), they 32

attracted the interest of researcher and, more recently, industries by their singularities. Isolated 33

in the deep ocean, far from the sunlight and subsequent organic substrate, the primal energy 34

sources for the development of this luxuriant biosphere remain elusive in these extreme 35

environments rich in minerals. Since their discovery, many new metabolisms have been identified 36

based on organic or inorganic molecules. However, the driving force sustaining all biodiversity in 37

these environments is thought to be based on chemolithoautotrophy (Alain et al., 2004). Indeed, 38

unlike most ecosystems, deep-sea ecosystems are totally dark and microorganisms have adapted 39

to base their metabolism on lithoautotrophy using inorganic compounds as the energy source to 40

fix inorganic carbon sources. Primary colonizers of deep-sea hydrothermal vents are assumed to 41

be (hyper)thermophilic microbes developing near the hydrothermal fluid, as retrieved in young 42

hydrothermal chimneys. These first colonizers are affiliated to Archaea, such as Archaeoglobales, 43

Thermococcales, Methanococcales or Desulfurococcales, and to Bacteria from ε-proteobacteria 44

and Aquificales. (Huber et al., 2002, 2003; Nercessian et al., 2003; Takai et al., 2004). Recent 45

studies have also shown that hyperthermophilic Archaea, which count among the supposed first 46

colonizers, are able to quickly scan and fix onto surfaces to find the best conditions for growth 47

(Wirth et al., 2018). These hyperthermophilic microorganisms would fix inorganic carbon through 48

3 chemolithoautotrophic types of metabolism, using H2, H2S or CH4 as energy sources and oxidized 49

molecules such as oxygen, sulfur compounds, iron oxide or even nitrate as electron acceptors.

50

However, the discovery of the presence of an abiotic electric current across the chimney walls 51

(Yamamoto et al., 2017) prompted the hypothesis of a new type of microorganisms called 52

eletrotrophs having the capacity to use electrons from the abiotic electric current as an energy 53

source coupled with carbon fixation from CO2. This metabolism was identified a few years ago on 54

a mesophilic chemolithoautotrophic Fe(II)-oxidizing bacterium, Acidithiobacillus ferrooxidans 55

(Ishii et al., 2015). This strain was able to switch its source of energy from diffusible Fe²⁺ ions to 56

direct electron uptake from a polarized electrode. However, this feature has not yet been 57

demonstrated in deep-sea hydrothermal vent environments. Recent studies have shown the 58

exoelectrogenic ability of some hyperthermophilic microorganisms isolated from deep-sea 59

hydrothermal vents, belonging to Archaeoglobales and Thermococcales (Pillot et al., 2018, 2019;

60

Yilmazel et al., 2018), but no studies have been done on environmental samples potentially 61

harboring electrotrophic communities growing naturally with an electric current as sole energy 62

source.

63

Here, we investigate the potential presence of electrotrophic communities in deep-sea 64

hydrothermal vents capable of using electrons directly or indirectly from the abiotic current. In 65

this purpose, we mimic the conductive surface of the hydrothermal chimney in a cathodic 66

chamber of Microbial Electrochemical Systems (MES) with a polarized cathode to enrich the 67

potential electrotrophic communities inhabiting these extreme environments. The polarized 68

cathode served as the sole energy source, while CO2 bubbling served as sole carbon source. Three 69

electron acceptors were tested separately, i.e. nitrate, oxygen, and sulfate, to show the influence 70

of an electron acceptor on community taxonomic composition.

71

Results

72

Current consumption from electrotroph enrichments 73

4 Hydrothermal vents chimney samples were inoculated in MES filled with sterile mineral medium 74

and incubated at 80°C to enrich electrotrophic communities. In the latter, the electrode served as 75

the sole energy donor (cathode) and sparged CO2 as carbon source with three different electron 76

acceptors that were tested separately: (i) nitrate, (ii) sulfate and (iii) oxygen. The microbial 77

electrotrophic enrichment was monitored at lowest possible potentials. These potentials were 78

poised at -300mV/SHE in presence of oxygen and -590mV/SHE for both nitrate and sulfate, 79

respectively. For comparison, microbial growth was also monitored without any poised potential 80

during a month in the same conditions of incubation. Interestingly, in the latter condition, no 81

microbial growth occurred, supported by microscope and spectrophotometric observations (data 82

not shown). Moreover, no organic compounds were produced supported by the HPLC and NMR 83

measurements (data not shown).

84

When potential was poised, abiotic controls containing no inoculum displayed constant currents 85

of ≈0,016 A.m^-2 at -590 mV and ≈0,01 A.m^-2 at -300 mV/SHE. In both conditions, the potential 86

hydrogen production on the cathode by water electrolysis was quantified and was under the 87

detection threshold of the µGC (>0.001% of total gas), indicating a theoretical production lower 88

than 34 µM day-1 (data not shown), similar as previously reported at 25°C (Marshall et al., 2013).

89

In comparison, experiments with the chimney sample showed current consumptions increasing 90

shortly after inoculation (Fig. 1). Indeed, when subtracting abiotic current, the current 91

consumptions reached a stabilized maximum of 0.36 A.m^-2 on oxygen, 0.72 A.m^-2 on nitrate, and 92

up to 1.83 A.m^-2 on sulfate corresponding therefore to an increase of 36, 45 and 114-fold 93

compared to abiotic current, respectively. MES were autoclaved afterwards displaying decreased 94

currents that were similar to the values of abiotic controls with a stabilized current around ≈0,021 95

A.m^-2. 96

At the end of monitoring of current consumption, CycloVoltamograms (CV) were performed to 97

study reactions of oxidation and reduction that could occur in MES (Supporting Information Fig.

98

5 S1A). A first peak of reduction is observed at -0.295, -0.293 and -0.217 V vs SHE in presence of 99

nitrate, sulfate and oxygen as electron acceptor, respectively (Fig. S1B). A second peak is observed 100

at -0.639 and -0.502 V vs SHE on nitrate and sulfate, respectively. No redox peaks are detected in 101

the abiotic controls and freshly inoculated MES, hence indicating a lack of electron shuttles 102

brought with the inoculum (Fig. S1A).

103

Organic compounds production in liquid media 104

During enrichment of the electrotrophic communities, the production of organic compounds was 105

monitored in liquid media (Fig. 1). Interestingly, the glycerol, the pyruvate, and the acetate were 106

the dominant products released in all experiment runs. Glycerol increased slowly throughout the 107

experiments to reach a maximum of 0.47 mM on sulfate (day 11), 1.32 mM on oxygen (day 12) 108

and 2.32 mM on nitrate (day 19). Acetate accumulated in the medium to reach 0.33 mM on 109

oxygen (day 7), 0.75 mM on sulfate (day 13) and 1.40 mM on nitrate (day 19). Pyruvate was 110

produced after a few days of culture with an exponential curve, reaching a maximum of 1.32 mM 111

on oxygen (day 12), 2.39 mM on nitrate (day 9), and 3.94 mM on sulfate (day 11). Pyruvate varied, 112

afterwards, due probably to microbial consumption or thermal degradation... Coulombic 113

efficiency calculated on the last day of the experiment (Fig. 2) showed up to 71% (on nitrate), 114

89% (on oxygen) and 90% (on sulfate) of electrons consumed were converted to organic 115

compounds and released into the liquid media .The rest represents the share of electrons 116

retained in not accumulated compounds (Table S1) and in the organic matter constituting the 117

cells of the electrotrophic communities (estimated by qPCR to total between 10⁸ to 10¹⁰ 16S rRNA 118

gene copies per MES; Supporting Information Fig. S2).

119

Biodiversity of electrotrophic communities on different electron acceptors 120

Once current consumption reached a stabilized maximum, DNAs from the biofilm and from 121

planktonic cells in culture media were extracted and sequenced on the V4 region of the 16S rRNA 122

6 to study relative abundance of the biodiversity. Fig. 3 reports the taxonomic affiliation of the 123

OTUs obtained. The chimney fragment inoculum showed a rich biodiversity (Shannon index at 124

5.29 and Pielou’s index at 0.69), with 208 OTUs mainly affiliated to Bacteria (99.49% vs 0.51% of 125

Archaea) and more particularly to Proteobacteria from Vibrionales (34.8%), miscellaneous rare 126

Proteobacteria (>33%), Campylobacterales (8.3%), Thermales (7.1%), Aquificales (5.62%), and 127

Rhodobacterales (5.1%) . 128

Enrichments in MES showed less biodiversity on cathodes and liquid media, suggesting the 129

selective development of functional communities. The Shannon index values were 3.1 and 1.9 on 130

nitrate, 1.7 and 1.9 on sulfate, and 4.1 and 4.2 on oxygen, with fewer OTUs associated to 72 and 131

68 OTUs on nitrate, 39 and 53 on sulfate, and 94 and 102 on oxygen, on electrodes and liquid 132

media, respectively. The taxonomic composition of these communities showed a larger 133

proportion of Archaea, with 51% and 41.6% on nitrate, 96.8% and 96.3% on sulfate, and 7.6% and 134

3.4% on oxygen on the electrodes and liquid media, respectively. In presence of each electron 135

acceptor, the archaeal population on the electrode was mainly composed of Archaeoglobales and 136

Thermococcales at different relative abundances. The latter were present at 6.7% and 28.2% on 137

nitrate to 65.8% and 28.6% on sulfate and 3.8% and 3.6% on oxygen, respectively. Equivalent 138

proportions of Archaeoglobales and Thermococcales were retrieved in liquid media, at 1.0% and 139

2.5% on nitrate, 65.8% and 29.6% on sulfate and 0.3% and 2.7% on oxygen, respectively (Fig. 3).

140

The MiSeq Illumina results served to study only 290 bp of 16S rRNA and thus to affiliate 141

microorganisms confirmed at family level, but they can also provide some information on the 142

enriched genera. In an effort to obtain more information on the probable Archaeoglobales and 143

Thermococcales genus, we attempted a species-level identification through phylogenetic analysis.

144

The results are presented in Fig. 4 as a Maximum Likelihood phylogenetic tree. The dominant 145

OTUs on sulfate and oxygen were closest to Ferroglobus placidus and Archeoglobus fulgidus 146

97.61% whereas the dominant OTU on nitrate was affiliated to Geoglobus ahangari with an 147

7 identity of 98.63% of identity. The remaining part of the biodiversity was specific to each electron 148

acceptor used. Enrichment on nitrate showed 13.8% and 28% of Desulfurococcales and 46.2% and 149

56.5% of Thermales on the electrode and in liquid media, respectively. Among Thermales that 150

developed on the electrode, 30% were represented by a new taxon (OTU 14 in Fig. 4 and S3) 151

whose closest cultured species was Vulcanithermus mediatlanticus (90% similarity). On sulfate, 152

the remaining biodiversity represented less than 4% of the population but was mainly 153

represented by two particular OTUs. The first OTU, accounting for up to 2.4% and 0.8% of the 154

total population on the electrode and liquid media, respectively, was affiliated to a new 155

Euryarchaeota (OTU 10 in Fig. 4 and S3) whose closest cultured match was Methanothermus 156

fervidus strain DSM 2088, at 86% similarity. The second OTU accounted for 2.0% and 1.9% of the 157

biodiversity on the electrode and liquid media, respectively, and was affiliated to the new 158

Deinococcales (OTU 14 in Fig. 4 and S3) species, found mostly on the electrode in nitrate 159

enrichment. In the enrichment on oxygen, the communities are dominated by 36.6% and 30.2% of 160

Pseudomonadales (Pseudomonas sp.), 14% and 42.6% of Bacillales (Bacillus and Geobacillus sp.), 161

21.3% and 7.41% of Vibrionales (Photobacterium sp.), and 9.8% and 5.1% of Actinomycetales 162

(spread across 9 species) on the electrode and in liquid media, respectively, with the rest spread 163

across Proteobacteria orders.

164

The clustering of the dominant OTUs (at a threshold of 0.05% of total sequences) obtained 165

previously on the chimney sample and enrichments in MES showed a clear differentiation of 166

communities retrieved in each sample (Supporting Information Fig. S3). The Pearson method on 167

OTU distribution produced four clusters, one corresponding to the inoculum and the three others 168

to each electron acceptor. Indeed, only two OTUs (OTU 4 and 36) were clearly shared between 169

two different communities, one affiliated to Thermococcus spp. on nitrate and sulfate and one to 170

Ralstonia sp. on the chimney sample and on nitrate. It is surprising to observe a recurrence of this 171

last OTU which could be a contaminant specific to the extraction kit used (Salter et al., 2014). The 172

8 other 50 dominant OTUs were specific to one community, with 21 OTUs on oxygen, 4 on sulfate, 8 173

on nitrate, and 17 on the chimney sample. The electrotrophic communities, colonizing the 174

cathode, were therefore different depending on electron acceptor used and their concentration 175

was too low to be detected in the chimney sample.

176

Discussion

177

Archaeoglobales as systematic (electro)lithoautotrophs of the community 178

Herein, we evidenced the development of a microbial electrotrophic communities and metabolic 179

activity supported by current consumption (Fig.1), product production (Fig. 2), and qPCRs (Fig. S2) 180

suggesting that growth did occur from energy supplied by the cathode. The mechanism of energy 181

uptake from electrode is discussed since the discovery of biofilms growing on cathode, and little is 182

known, unlike electron transfer mechanism on anode. The two main hypotheses are the use of 183

similar direct electron transfer pathway as on the anode, or the use of molecular H2, produced by 184

water electrolysis, as electron mediator to the cell. In both cases, our study is the first to show the 185

possibility of growth of biofilm from environments harboring natural electric current in absence of 186

organic substrates. To discuss further on the putative mechanism, it is necessary to have a look on 187

the conditions for water electrolysis. The potential for water reduction into hydrogen at 80°C, ph7, 188

1 atm was calculated at −0.490 V vs SHE in pure water. The real operational reduction potentials is 189

expected to be much more lower than the theoretical value due to internal resistances (from 190

electrical connections, electrolyte, ionic membrane etc.) (Lim et al., 2017). Moreover, 191

overpotentials are expected with carbon electrodes. The decrease of this potential explains the 192

absence of hydrogen measured in our conditions. A screening of potential in abiotic conditions 193

confirmed the increase of current consumption and H2 production only at potential lower of -0.6V 194

vs SHE (Fig S1) Moreover several pieces of evidence indicate that direct electron transfer may have 195

mainly participated in the development of biofilms: the growth of the similar dependent sulfate 196

biodiversity with the cathode poised at -300 mV vs SHE (Supporting Information Fig. S4) without H2

197

9 production, the expression of catalytic waves observed by CV with midpoint potentials between - 198

0.217 V to -0.639 V and the lack of similar peaks with abiotic or fresh inoculated media (Supporting 199

Information Fig. S1), biofilm formation on the electrode (as on nitrate, Supporting Information Fig.

200

S5), delayed production of glycerol, pyruvate and acetate (Fig. 1) fixing between 267 to 1596 201

Coulombs.day^-1 (organic consumption deduced), and the recovery of electrons in all three products 202

(Fig. 2), that largely exceeds the maximum theoretical abiotic generation of hydrogen (~3 C.day^-1) 203

by 90 to 530-fold. Thus, we can assume that the biofilm growth was largely ensured by a significant 204

part of a direct transfer of electrons from the cathode demonstrating the presence of 205

electrolithoautotroph microorganisms.

206

Taxonomic analysis of the enriched microbial communities at the end of the experiments showed 207

the systematic presence on cathodes of Archaeoglobales (Fig. 3 and S3), whatever the electron 208

acceptors used. The species belonging to Archaeoglobales order were the only enriched species in 209

all conditions and the only known to have an autotrophic metabolism (except for Archaeoglobus 210

profundus and A. infectus which are obligate heterotrophs). The Archaeoglobales order is 211

composed of three genera: Archaeoglobus, Geoglobus, and Ferroglobus. All are hyperthermophilic 212

obligate anaerobes with diverse metabolisms, including heterotrophy or chemolithoautotrophy.

213

Terminal electron acceptors used by this order include sulfate, nitrate, poorly crystalline Fe (III) 214

oxide, or sulfur oxyanions (Brileya and Reysenbach, 2014). Autotrophic growth in the 215

Archaeoglobales order is ensured mainly through H2 as energy source and requires both branches 216

of the reductive acetyl-CoA/Wood-Ljungdahl pathway for CO2 fixation (Vorholt et al., 1997).

217

Moreover, Archaeoglobus fulgidus has been recently shown to grow on iron by directly snatching 218

electrons under carbon starvation during corrosion process (Jia et al., 2018). Furthermore, 219

Ferroglobus and Geoglobus species were shown to be exoelectrogens in pure culture in a microbial 220

electrosynthesis cell (Yilmazel et al., 2018) and were enriched during a study within a microbial 221

electrolysis cell (Pillot et al., 2018, 2019). Interestingly, some studies have shown that Geobacter 222

10 species are capable of bidirectional electron transfer using the same mechanism (Pous et al., 2016).

223

Hence, Archaeoglobales that have been shown already as exoelectrogens (Yilmazel et al., 2018) 224

could also be electrotrophs. It is not known how Archaea carry out exogenous electron transfer. As 225

previously discussed, the absence of H2 production and the increasing current consumptions over 226

time suggest direct electron uptake from members of the communities developing on the 227

electrode, as for Acidithiobacillus ferroxidans (Ishii et al., 2015). Moreover, the qPCR (Supporting 228

Information Fig. S2) and MiSeq data (Fig. 3) highlighted a strong correlation between current 229

consumption and density of Archaeoglobales on the electrode (R²=0.962). In the condition with 230

sulfate as electron acceptor, the proportion of Archaeoglobales represented 65.8% of total 231

biodiversity providing 1.83 A.m^-2 of current consumption, compared to only 6.7% in the nitrate 232

condition and 3.8% in the oxygen condition for 0.72 and 0.36 A.m^-2 of current consumption, 233

respectively. Moreover, the majority of OTUs were affiliated to three Archaeoglobaceae genera, 234

mainly Archaeoglobus spp. and Ferroglobus spp. on sulfate and oxygen and Geoglobus sp. on 235

nitrate. Some Archaeoglobus are known to show anaerobic sulfate-reducing metabolism while 236

Ferroglobus spp. are not. Geoglobus sp. has never been described to perform nitrate reduction so 237

far, but it does harbor genes of nitrate- and nitrite-reductase-like proteins (Manzella et al., 2015).

238

The OTUs were related to some Archaeoglobales strain with 95-98% identities. Thus, we assume 239

that in our conditions, new specific electrotrophic metabolisms or new electrolithoautotrophic 240

Archaeoglobaceae species were enriched on the cathode. Moreover, a member of a new 241

phylogenetic group of Archaea was enriched up to 2.4% of total biodiversity on sulfate (OTU10 in 242

Fig. 4). While its metabolism is still unknown, we suggest that isolation of this electrotrophic 243

archaea in MES could enable the identification of a new archaeal phylogenetic group based on 244

electrotrophy.

245

The growth of Archaeolgobales species in presence of oxygen is a surprising finding.

246

Archaeoglobales have a strictly anaerobic metabolism, and the reductive acetyl-CoA pathway is 247

11 very sensitive to the presence of oxygen (Fuchs, 2011). This can be firstly explained by the low 248

solubility of oxygen at 80°C combined with the electrochemically oxygen reduction on electrode 249

in controls (data not shown). It hence results in an low oxygen or oxygen-free environment within 250

the carbon cloth mesh for anaerobic development of microorganisms into a protective biofilm 251

(Hamilton, 1987). This observation was also supported by the near absence of Archaeoglobales in 252

the liquid media (Fig. 3). In absence of other electron acceptors, some Archaeoglobales perform 253

carboxydotrophic metabolism to grow from CO, as demonstrated for Archaeoglobus fulgidus 254

(Sokolova and Lebedinsky, 2013; Hocking et al., 2015). This fermentative CO metabolism leads to 255

the production of acetate and transient accumulation of formate via the Wood-Ljungdahl 256

pathway, but no net ATP is really produced (Henstra et al., 2007). The energy conservation 257

through this metabolism in Archaeoglobus fulgidus is still poorly understood (Hocking et al., 258

2015). A second hypothesis concerns direct interspecies electron transfer (DIET) (Kato et al., 259

2012; Lovley, 2017), with Archaeoglobales transferring electrons to another microorganism as an 260

electron acceptor. Research into DIET is in its early stages, and further investigations are required 261

to better understand the diversity of microorganisms and the mechanism of carbon and electron 262

flows in anaerobic environments (Lovley, 2017) such as hydrothermal ecosystems.

263

Electrosynthesis of organic compounds 264

The pyruvate, the glycerol and the acetate accumulated, while another set of compounds that 265

appear transiently were essentially detectable in the first few days of biofilm growth (Supporting 266

Information Table S1). They included amino acids (threonine, alanine) and volatile fatty acids 267

(formate, succinate, lactate, acetoacetate, 3-hydroxyisovalerate) whose concentrations did not 268

exceed one micromole. Despite their thermostability, this transient production suggests they 269

were used by microbial communities developing on the electrode in interaction with the primary 270

producers during enrichment.

271

12 On the other hand, in presence of nitrate, sulfate and oxygen as electron acceptors, the liquid 272

media accumulated three main organic products acetate, glycerol, and pyruvate (Fig. 1).

273

Coulombic efficiency calculations (Fig. 2) showed that redox levels of the carbon-products 274

represented 71%–90% of electrons consumed, and only about 10%–30% of net electrons 275

consumed by electrotrophs during growth was used directly for biomass or transferred to an 276

electron acceptor. This concurs with the energy yield from the Wood-Ljungdahl pathway of 277

Archaeoglobales, with only 5% of carbon flux directed to the production of biomass and the other 278

95% diverted to the production of small organic end-products excreted from the cell (Fast and 279

Papoutsakis, 2012).

280

However, the production of pyruvate and glycerol warrants further analysis. Pyruvate is normally 281

a central intermediate of CO2 uptake by the reducing route of the acetyl-CoA/WL pathway (Berg 282

et al., 2010). It can be used to drive the anabolic reactions needed for biosynthesis of cellular 283

constituents. Theoretically, the only explanation for improved production and accumulation of 284

pyruvate (up to 5 mM in the liquid media of sulfate experiment) would be that pyruvate-using 285

enzymes were inhibited or that pyruvate influx exceeded its conversion rate. Here we could 286

suggest that in-cell electron over-feeding at the cathode leads to significant production of 287

pyruvate. Indeed, in a physiological context, the production of pyruvate from acetyl-CoA via 288

pyruvate synthase requires the oxidation of reduced ferredoxins for CO2 fixation (Furdui and 289

Ragsdale, 2000). The continuous electron uptake from the cathode would lead to a significant 290

reduction in electron carriers (including ferredoxins, flavins, cytochromes, and/or nicotinamides), 291

thus forcing the electrotrophic microbial community to produce pyruvate as a redox sink.

292

In the same context of pyruvate production, glycerol is produced by reduction of 293

dihydroxyacetone phosphate a glycolytic intermediate, to glycerol 3-phosphate (G3P) followed by 294

dephosphorylation of G3P to glycerol. In some yeasts, glycerol production is essential for 295

osmoadaptation but equally for regulating the NADH surplus during anaerobic growth (Björkqvist 296

13 et al., 1997). A similar mechanism may operate in our conditions for the probable excess of NADH 297

pool due to the electrode poised at -590 mV vs SHE, which would explain the accumulation of 298

glycerol found in our experiments.

299

In an ecophysiological context, a similar pyruvate and glycerol production could occur on 300

hydrothermal chimney walls into which electric current propagates (Yamamoto et al., 2017). The 301

electrotroph biofilms would continually receive electrons, leading to the excess of intracellular 302

reducing power that would be counterbalanced by the overproduction of glycerol and pyruvate.

303

Moreover, glycerol is an essential compound in the synthesis of membrane lipids in Archaea and 304

probably also in biofilm formation and osmoadaptability (Desai et al., 2013; Shemesh and Chai, 305

2013). Pyruvate unites several key metabolic processes, such as its conversion into carbohydrates, 306

fatty acids or some amino acids. Furthermore, these products can serve as carbon and energy 307

sources for heterotrophic microorganisms or for fermentation. In our experiments, pyruvate and 308

glycerol concentrations varied over time, suggesting they were being consumed by heterotrophic 309

microorganisms. Acetate production would thus result from the fermentation of pyruvate or 310

other compounds produced by electrotrophic Archaeoglobales.

311

Enrichment of rich heterotrophic biodiversity from electrotrophic Archaeoglobales 312

community 313

During our enrichment experiments, the development of effective and specific biodiversity was 314

dependent on the electron acceptors used (Fig. 3). Heatmap analyses (Supporting Information Fig.

315

3) showed four distinct communities for the three electron acceptors and the initial inoculum. Thus, 316

at the lower taxonomic level of the biodiversity analysis, most OTUs are not shared between each 317

enrichment, except for one OTU of Thermococcales that was shared between the nitrate and sulfate 318

experiments. This suggests a real specificity of the communities and a specific evolution or 319

adaptation of the members of the shared phyla to the different electron acceptors available in the 320

environment. However, the various enrichments also showed the presence of Thermococcales 321

14 regardless of the electron acceptors used, thus demonstrating a strong interaction between 322

Thermococcales, heterotrophs, and Archaeoglobales, the only autotrophs. In a previous study, 323

enrichments on the anode of a microbial electrolysis cell showed a similar tendency, with 324

Archaeoglobales strongly correlated to Thermococcales (Pillot et al., 2018, 2019). Moreover, 325

members of these two groups have frequently been found together in various hydrothermal sites 326

on the surface of the Earth (Corre et al., 2001; Nercessian et al., 2003; Takai et al., 2004; Jaeschke 327

et al., 2012), where they are considered as potential primary colonizers of their environments (33–

328

36). This could point to a co-evolution and metabolic adaptation of these microorganisms to their 329

unstable environmental conditions in hydrothermal settings. After Thermococcales, the rest of the 330

heterotrophic biodiversity was specific to each electron acceptor.

331

On nitrate, two additional phylogenetic groups were retrieved: Desulfurococcales and Thermales.

332

OTUs of Desulfurococcales are mainly affiliated to Thermodiscus or Aeropyrum species, which are 333

hyperthermophilic and heterotrophic Crenarchaeota growing by fermentation of complex organic 334

compounds or sulfur/oxygen reduction (Huber and Stetter, 2015).

335

Concerning Thermales, a new taxon was enriched on cathode and only affiliated with 90 % 336

similarity to Vulcanithermus mediatlanticus. On sulfate, beside the large majority of 337

Archaeoglobales and Thermococcales (up to 94%–96%), this new taxon of Thermales (OTU 14, Fig.

338

S3) has also been enriched on the cathode, representing 2% of total biodiversity. Thermales are 339

thermophilic (30°C–80°C) and heterotrophic bacteria whose only four genera (Marinithermus, 340

Oceanithermus, Rhabdothermus, and Vulcanithermus) are all retrieved in marine hydrothermal 341

systems. They are known to be aerobic or microaerophilic. Some strains grow anaerobically with 342

several inorganic electron acceptors such as nitrate, nitrite, Fe (III) and elemental sulfur 343

(Albuquerque and Costa, 2014). All of the species Thermales can utilize the pyruvate as carbon 344

and energy source. The produced pyruvate would be a substrate of choice for this new taxon 345

which would use the sulfate and nitrate as electron acceptors.

346

15 Pseudomonadales and Bacillales were found in the oxygen experiment. Most Pseudomonas are 347

known to be aerobic and mesophilic bacteria, with a few thermophilic species, including the 348

autotrophic Pseudomonas thermocarboxydovorans that grows at up to 65°C (Lyons et al., 1984;

349

Palleroni, 2015). There have already been reports of mesophilic Pseudomonas species growing in 350

thermophilic conditions in composting environments (Droffner et al., 1995). Moreover, some 351

Pseudomonas sp. are known to be electroactive in microbial fuel cells, through long-distance 352

extracellular electron transport (Shen et al., 2014; Maruthupandy et al., 2015; Lai et al., 2016), 353

and were dominant on the cathodes of a benthic microbial fuel cell on a deep-ocean cold seep 354

(Reimers et al., 2006). In Bacillales, the Geobacillus spp. and some Bacillus sp. are known to be 355

mainly (hyper)thermophilic aerobic and heterotrophic Firmicutes (Vos, 2015).

356

Hydrothermal electric current: a new energy source for the development of primary 357

producers 358

The presence of so many heterotrophs in an initially autotrophic condition points to the 359

hypothesis of a trophic relationship inside the electrotrophic community (Fig. 5). This suggests 360

that the only autotrophs retrieved in all communities, the Archaeoglobales, might be the first 361

colonizer of the electrode, using CO2 as carbon source and cathode as energy source. Studies have 362

shown how modeling and field observations can be usefully combined to describe the relationship 363

between chemical energy conditions and metabolic interactions within microbial communities 364

(Lin et al., 2016; Dahle et al., 2018). However, the models predicted low abundances of 365

Archaeoglobales (<0.04%) whereas on-field detection found abundances of more than 40% in the 366

inner section of the studied hydrothermal chimney (Dahle et al., 2018). Indeed, in these models, 367

the predicted H2 concentration, based on observations, would be too low to support the growth 368

of hydrogenotrophic or methanogenic species (Lin et al., 2016). The authors concluded on a 369

probable H2 syntrophy, with hydrogen being produced by heterotrophic microorganisms such as 370

fermentative Thermococcales species. Our study is the first evidence of the development of 371

16 hyperthermophilic electrotrophic/heterotrophic communities directly enriched from the natural 372

environment known to harbor natural electric current as a potential energy source. We can thus 373

conclude that this kind of electrolithoautotrophic metabolism is highly likely in deep-sea 374

hydrothermal ecosystems, which raises the question of the importance of this metabolism in the 375

primary colonization of hydrothermal vents. The hydrothermal electric current could make up for 376

the lack of H2 normally needed to sustain the growth of hydrogenotrophic microorganisms.

377

Indeed, the constant electron supply on the surface of a conductive chimney allows a new energy 378

source and long-range transfer between the electron donor (represented here by reduced 379

molecules such as H2S electrochemically oxidized on the inner surface of the chimney wall) and 380

the electron acceptor (O2, sulfur compounds, nitrate, metals) present all over the external surface 381

of the chimney. This electrical current would thus allow primary colonizers to grow not just on all 382

the surface but also in the chimney structure. These primary colonizers would release organic 383

compounds used by the heterotrophic community for growth, as observed with the successive 384

production and consumption of organic compounds in our experiments. Moreover, migrating out 385

to larger potential growth surface would help to meet a wider range of physiological conditions 386

through pH, temperature and oxidoreduction gradients. This allows a wider diversity of growth 387

patterns than through chemolithoautotrophy, which is restricted to unstable and limited contact 388

zones between reduced compounds (H2, H2S) in the hydrothermal fluid and electron acceptors 389

around the hydrothermal chimneys, often precipitating together.

390

Conclusion

391

Taken together, the results found in this study converge into evidence of the ability of indigenous 392

microorganisms from deep hydrothermal vents to grow using electric current and CO2. This ability 393

seems to be spread across diverse phylogenetic groups and to be coupled with diverse electron 394

acceptors. Through their electro-litho-auto-trophic metabolism, Archaeoglobaceae strains 395

produce and release organic compounds into their close environment, allowing the growth of 396

17 heterotrophic microorganisms, and ultimately enabling more and more diversity to develop over 397

time. This metabolism could be one of the primary energies for the colonization of deep-sea 398

hydrothermal chimneys and the development of a complex trophic network driving sustainable 399

biodiversity. A similar mechanism could have occurred during the Hadean, allowing the 400

emergence of life in hydrothermal environments by constant electron influx to the first proto- 401

cells.

402

Experimental procedures

403

Sample collection and preparation 404

A hydrothermal chimney sample was collected on the acidic and iron-rich Capelinhos site on the 405

Lucky Strike hydrothermal field (37°17.0'N, MAR) during the MoMARsat cruise in 2014 406

(http://dx.doi.org/10.17600/14000300) led by IFREMER (France) onboard R/V Pourquoi Pas?

407

(Sarradin and Cannat, 2014). The sample (PL583-8) was collected by breaking off a piece of a high- 408

temperature active black smoker using the submersible’s robotic arm, and bringing it back to the 409

surface in a decontaminated insulated box (http://video.ifremer.fr/video?id=9415). Onboard, 410

chimney fragments were anaerobically crushed in an anaerobic chamber under H2:N2 (2.5:97.5) 411

atmosphere (La Calhene, France), placed in flasks under anaerobic conditions (anoxic seawater at 412

pH 7 with 0.5 mg L^-1 of Na2S and N2:H2:CO2 (90:5:5) gas atmosphere), and stored at 4°C.

413

Prior to our experiments, pieces of the hydrothermal chimney were removed from the sulfidic 414

seawater flask, crushed with a sterile mortar and pestle in an anaerobic chamber (Coy 415

Laboratories, Grass Lake, MI), and distributed into anaerobic tubes for use in the various 416

experiments.

417

Electrotrophic enrichment on nitrate, sulfate, and oxygen 418

MES were filled with 1.5 L of an amended sterile mineral medium as previously described (Pillot 419

et al., 2018) without yeast extract, and set at 80°C and pH 6.0 throughout on-platform monitoring.

420

The electrode (cathode) composed of 20 cm² of carbon cloth was poised at the lowest potential 421

18 before initiation of abiotic current consumption (Supporting Information Fig S6) using SP-240 422

potentiostats and EC-Lab software (BioLogic, France). We thus used a potential of -590 mV vs( in 423

the nitrate and sulfate experiments and -300 mV vs SHE in the oxygen experiment. A similar 424

experiment at -300 mV vs SHE has been initiated in presence of sulfate (see SI Fig. S4) to confirm 425

the growth of electrolithoautotroph microorganisms without any H2 production possible. The 426

electrode poised as cathode served as the sole electron donor for electrotroph growth. For the 427

nitrate experiment, one system was supplemented with 4 mM of sodium nitrate. For the sulfate 428

experiment, a second system was supplemented with 10 mM of sodium sulfate, and the cathodic 429

chambers were sparged with N2:CO2 (90:10, 100 mL/min). For the oxygen experiment, a third 430

system was sparged with N2:CO2:O2 (80:10:10, 100 mL/min) with initially 10% oxygen as electron 431

acceptor. All three systems were inoculated with 8 g of the crushed chimney (~0.5% (w/v)).

432

Current consumption was monitored via the chronoamperometry method with current density 433

and readings were taken every 10 s. An abiotic control without inoculation showed no current 434

consumption during the same experiment period. CycloVoltammograms (scan rate: 20 mV/s) 435

were analyzed using QSoas software (version 2.1). Coulombic efficiencies where calculates with 436

the following equation:

437

𝐶𝐸 (%) =F ∙ n_e∙ ∆[P] ∙ V_catholyte

∫ I(t) ∙ dt_t^t

0

∙ 100 438

I(t): current consumed between t0 and t (A) 439

F: Faraday constant 440

ne: number of moles of electrons presents per mole of product (mol) 441

∆[P]: variation of the concentration of organic product between t0 and t (mol.L^-1) 442

Vcatholyte: volume of catholyte (L) 443

Identification and quantification of organic compound production 444

19 To identify and quantify the production of organic compounds from the biofilm, samples of liquid 445

media were collected at the beginning and at the end of the experiment and analyzed by ¹H NMR 446

spectroscopy. For this, 400 µL of each culture medium, were added to 200 L of PBS solution 447

prepared in D2O (NaCl, 140 mM; KCl, 2.7 mM; KH2PO4, 1.5 mM; Na2HPO4, 8.1 mM, pH 7.4) 448

supplemented with 0.5 mmol L^-1 of trimethylsilylpropionic acid-d4 (TSP) as NMR reference. All the 449

1D ¹H NMR experiments were carried out at 300 K on a Bruker Avance spectrometer (Bruker, 450

BioSpin Corporation, France) operating at 600 MHz for the ¹H frequency and equipped with a 5- 451

mm BBFO probe.

452

Spectra were recorded using the 1D nuclear Overhauser effect spectroscopy pulse sequence (Trd- 453

90°-T1-90°-tm-90°-Taq) with a relaxation delay (Trd) of 12.5 s, a mixing time (tm) of 100 ms, and a 454

T1 of 4 μs. The sequence enables optimal suppression of the water signal that dominates the 455

spectrum. We collected 128 free induction decays (FID) of 65,536 datapoints using a spectral 456

width of 12 kHz and an acquisition time of 2.72 s. For all spectra, FIDs were multiplied by an 457

exponential weighting function corresponding to a line broadening of 0.3 Hz and zero-filled before 458

Fourier transformation. NMR spectra were manually phased using Topspin 3.5 software (Bruker 459

Biospin Corporation, France) and automatically baseline-corrected and referenced to the TSP 460

signal (δ = -0.015 ppm) using Chenomx NMR suite v7.5 software (Chenomx Inc., Canada). A 0.3 Hz 461

line-broadening apodization was applied prior to spectral analysis, and ¹H-¹H TOCSY (Bax and 462

Davis, 1985) and ¹H-¹³C HSQC (Schleucher et al., 1994) experiments were recorded on selected 463

samples to identify the detected metabolites. Quantification of identified metabolites was done 464

using Chenomx NMR suite v7.5 software (Chenomx Inc., Canada) using the TSP signal as the 465

internal standard.

466

Biodiversity analysis 467

Taxonomic affiliation was carried out according to (Zhang et al., 2016). DNA was extracted from 1 468

g of the crushed chimney and, at the end of each culture period, from scrapings of half of the WE 469

20 and from centrifuged pellets of 50 mL of spent media. The DNA extraction was carried out using 470

the MoBio PowerSoil DNA isolation kit (Carlsbad, CA). The V4 region of the 16S rRNA gene was 471

amplified using the universal primers 515F (5′-GTG CCA GCM GCC GCG GTA A-3′) and 806R (5′- 472

GGA CTA CNN GGG TAT CTA AT-3′) (Bates et al., 2011) with Taq&Load MasterMix (Promega). PCR 473

reactions, qPCR, amplicon sequencing and taxonomic affiliation were carried as previously 474

described (Pillot et al., 2018). The qPCR results were expressed in copies number of 16s rRNA 475

gene per gram of crushed chimney, per milliliter of liquid media or per cm² of surface of the 476

electrode. To analyze alpha diversity, the OTU tables were rarefied to a sampling depth of 9410 477

sequences per library, and three metrics were calculated: the richness component, represented 478

by number of OTUs observed, the Shannon index, representing total biodiversity, and the 479

evenness index (Pielou’s index), which measures distribution of individuals within species 480

independently of species richness. Rarefaction curves (Supporting Information Fig. S7) for each 481

enrichment approached an asymptote, suggesting that the sequencing depths were sufficient to 482

capture overall microbial diversity in the studied samples. The phylogenetic tree was obtained 483

with MEGA software v10.0.5 with the MUSCLE clustering algorithm and the Maximum Likelihood 484

Tree Test with a Bootstrap method (2500 replications). The heatmap was obtained using RStudio 485

software v3. The raw sequences for all samples can be found in the European Nucleotide Archive 486

(accession number: PRJEB35427).

487

Acknowledgments

488

This work received financial support from the CNRS-sponsored national interdisciplinary research 489

program (PEPS-ExoMod 2016). The authors thank Céline Rommevaux and Françoise Lesongeur for 490

taking samples during the MOMARSAT 2014 cruise, the MIM platform (MIO, France) for providing 491

access to their confocal microscopy facility, and the GeT-PlaGe platform (GenoToul, France) for 492

help with DNA sequencing. The project leading to this publication received European FEDER 493

funding under project “1166-39417. The authors declare no conflicts of interest.

494

21 495

22

References

496

Alain, K., Zbinden, M., Bris, N.L., Lesongeur, F., Quérellou, J., Gaill, F., and Cambon‐Bonavita, 497

M.-A. (2004) Early steps in microbial colonization processes at deep-sea hydrothermal 498

vents. Environ Microbiol 6: 227–241.

499

Albuquerque, L. and Costa, M.S. da (2014) The Family Thermaceae. In The Prokaryotes. Springer, 500

Berlin, Heidelberg, pp. 955–987.

501

Bates, S.T., Berg-Lyons, D., Caporaso, J.G., Walters, W.A., Knight, R., and Fierer, N. (2011) 502

Examining the global distribution of dominant archaeal populations in soil. ISME J 5: 908–

503

917.

504

Bax, A. and Davis, D.G. (1985) MLEV-17-based two-dimensional homonuclear magnetization 505

transfer spectroscopy. J Magn Reson 1969 65: 355–360.

506

Berg, I.A., Kockelkorn, D., Ramos-Vera, W.H., Say, R.F., Zarzycki, J., Hügler, M., et al. (2010) 507

Autotrophic carbon fixation in archaea. Nat Rev Microbiol 8: 447–460.

508

Björkqvist, S., Ansell, R., Adler, L., and Lidén, G. (1997) Physiological response to anaerobicity of 509

glycerol-3-phosphate dehydrogenase mutants of Saccharomyces cerevisiae. Appl Env 510

Microbiol 63: 128–132.

511

Brileya, K. and Reysenbach, A.-L. (2014) The Class Archaeoglobi. In The Prokaryotes. Springer, 512

Berlin, Heidelberg, pp. 15–23.

513

Corliss, J.B. and Ballard, R.D. (1977) Oasis of life in the cold abyss. Nat Geogr Mag 152: 440–453.

514

Corre, E., Reysenbach, A.-L., and Prieur, D. (2001) ɛ-Proteobacterial diversity from a deep-sea 515

hydrothermal vent on the Mid-Atlantic Ridge. FEMS Microbiol Lett 205: 329–335.

516

Dahle, H., Le Moine Bauer, S., Baumberger, T., Stokke, R., Pedersen, R.B., Thorseth, I.H., and 517

Steen, I.H. (2018) Energy landscapes in hydrothermal chimneys shape distributions of 518

primary producers. Front Microbiol 9:.

519

Desai, J.V., Bruno, V.M., Ganguly, S., Stamper, R.J., Mitchell, K.F., Solis, N., et al. (2013) 520

Regulatory Role of Glycerol in Candida albicans Biofilm Formation. mBio 4:.

521

Droffner, M.L., Brinton, W.F., and Evans, E. (1995) Evidence for the prominence of well 522

characterized mesophilic bacteria in thermophilic (50–70°C) composting environments.

523

Biomass Bioenergy 8: 191–195.

524

Fast, A.G. and Papoutsakis, E.T. (2012) Stoichiometric and energetic analyses of non-photosynthetic 525

CO2-fixation pathways to support synthetic biology strategies for production of fuels and 526

chemicals. Curr Opin Chem Eng 1: 380–395.

527

Fuchs, G. (2011) Alternative Pathways of Carbon Dioxide Fixation: Insights into the Early Evolution 528

of Life? Annu Rev Microbiol 65: 631–658.

529

Furdui, C. and Ragsdale, S.W. (2000) The role of pyruvate ferredoxin oxidoreductase in pyruvate 530

synthesis during autotrophic growth by the Wood-Ljungdahl pathway. J Biol Chem 275:

531

28494–28499.

532

Hamilton, W.A. (1987) Biofilms : Microbial interactions and metabolic activities. Ecol Microb 533

Communities 361–385.

534

Henstra, A.M., Dijkema, C., and Stams, A.J.M. (2007) Archaeoglobus fulgidus couples CO oxidation 535

to sulfate reduction and acetogenesis with transient formate accumulation: The CO 536

metabolism of A. fulgidus. Environ Microbiol 9: 1836–1841.

537

Hocking, W.P., Roalkvam, I., Magnussen, C., Stokke, R., and Steen, I.H. (2015) Assessment of the 538

Carbon Monoxide Metabolism of the Hyperthermophilic Sulfate-Reducing Archaeon 539

Archaeoglobus fulgidus VC-16 by Comparative Transcriptome Analyses. Archaea 2015: 1–

540

12.

541

Huber, H. and Stetter, K.O. (2015) Desulfurococcales ord. nov. In Bergey’s Manual of Systematics 542

of Archaea and Bacteria. American Cancer Society, pp. 1–2.

543

Huber, J.A., Butterfield, D.A., and Baross, J.A. (2003) Bacterial diversity in a subseafloor habitat 544

following a deep-sea volcanic eruption. FEMS Microbiol Ecol 43: 393–409.

545

Huber, J.A., Butterfield, D.A., and Baross, J.A. (2002) Temporal changes in archaeal diversity and 546

chemistry in a mid-ocean ridge subseafloor habitat. Appl Environ Microbiol 68: 1585–1594.

547

23 Ishii, T., Kawaichi, S., Nakagawa, H., Hashimoto, K., and Nakamura, R. (2015) From 548

chemolithoautotrophs to electrolithoautotrophs: CO2 fixation by Fe(II)-oxidizing bacteria 549

coupled with direct uptake of electrons from solid electron sources. Front Microbiol 6:.

550

Jaeschke, A., Jørgensen, S.L., Bernasconi, S.M., Pedersen, R.B., Thorseth, I.H., and Früh‐Green, 551

G.L. (2012) Microbial diversity of Loki’s Castle black smokers at the Arctic Mid-Ocean 552

Ridge. Geobiology 10: 548–561.

553

Jia, R., Yang, D., Xu, D., and Gu, T. (2018) Carbon steel biocorrosion at 80 °C by a thermophilic 554

sulfate reducing archaeon biofilm provides evidence for its utilization of elemental iron as 555

electron donor through extracellular electron transfer. Corros Sci 145: 47–54.

556

Kato, S., Hashimoto, K., and Watanabe, K. (2012) Microbial interspecies electron transfer via 557

electric currents through conductive minerals. Proc Natl Acad Sci 109: 10042–10046.

558

Lai, B., Yu, S., Bernhardt, P.V., Rabaey, K., Virdis, B., and Krömer, J.O. (2016) Anoxic metabolism 559

and biochemical production in Pseudomonas putida F1 driven by a bioelectrochemical 560

system. Biotechnol Biofuels 9: 39.

561

Lim, S.S., Yu, E.H., Daud, W.R.W., Kim, B.H., and Scott, K. (2017) Bioanode as a limiting factor 562

to biocathode performance in microbial electrolysis cells. Bioresour Technol 238: 313–324.

563

Lin, T.J., Ver Eecke, H.C., Breves, E.A., Dyar, M.D., Jamieson, J.W., Hannington, M.D., et al.

564

(2016) Linkages between mineralogy, fluid chemistry, and microbial communities within 565

hydrothermal chimneys from the Endeavour Segment, Juan de Fuca Ridge:

566

GEOMICROBIOLOGY OF HYDROTHERMAL CHIMNEYS. Geochem Geophys 567

Geosystems 17: 300–323.

568

Lovley, D.R. (2017) Syntrophy goes electric: Direct interspecies electron transfer. Annu Rev 569

Microbiol 71: 643–664.

570

Lyons, C.M., Justin, P., Colby, J., and Williams, E. (1984) Isolation, characterization and autotrophic 571

metabolism of a moderately thermophilic carboxydobacterium, Pseudomonas 572

thermocarboxydovorans sp. nov. Microbiology 130: 1097–1105.

573

Manzella, M.P., Holmes, D.E., Rocheleau, J.M., Chung, A., Reguera, G., and Kashefi, K. (2015) The 574

complete genome sequence and emendation of the hyperthermophilic, obligate iron-reducing 575

archaeon “Geoglobus ahangari” strain 234T. Stand Genomic Sci 10: 77.

576

Marshall, C.W., Ross, D.E., Fichot, E.B., Norman, R.S., and May, H.D. (2013) Long-term Operation 577

of Microbial Electrosynthesis Systems Improves Acetate Production by Autotrophic 578

Microbiomes. Environ Sci Technol 47: 6023–6029.

579

Maruthupandy, M., Anand, M., Maduraiveeran, G., Beevi, A.S.H., and Priya, R.J. (2015) Electrical 580

conductivity measurements of bacterial nanowires from Pseudomonas aeruginosa. Adv Nat 581

Sci Nanosci Nanotechnol 6: 045007.

582

Nercessian, O., Reysenbach, A.-L., Prieur, D., and Jeanthon, C. (2003) Archaeal diversity associated 583

with in situ samplers deployed on hydrothermal vents on the East Pacific Rise (13°N).

584

Environ Microbiol 5: 492–502.

585

Palleroni, N.J. (2015) Pseudomonas. In Bergey’s Manual of Systematics of Archaea and Bacteria.

586

American Cancer Society, pp. 1–1.

587

Pillot, G., Davidson, S., Auria, R., Combet-Blanc, Y., Godfroy, A., and Liebgott, P.-P. (2019) 588

Production of Current by Syntrophy Between Exoelectrogenic and Fermentative 589

Hyperthermophilic Microorganisms in Heterotrophic Biofilm from a Deep-Sea 590

Hydrothermal Chimney. Microb Ecol.

591

Pillot, G., Frouin, E., Pasero, E., Godfroy, A., Combet-Blanc, Y., Davidson, S., and Liebgott, P.-P.

592

(2018) Specific enrichment of hyperthermophilic electroactive Archaea from deep-sea 593

hydrothermal vent on electrically conductive support. Bioresour Technol 259: 304–311.

594

Pous, N., Carmona-Martínez, A.A., Vilajeliu-Pons, A., Fiset, E., Bañeras, L., Trably, E., et al. (2016) 595

Bidirectional microbial electron transfer: Switching an acetate oxidizing biofilm to nitrate 596

reducing conditions. Biosens Bioelectron 75: 352–358.

597

Reimers, C.E., Girguis, P., Stecher, H.A., Tender, L.M., Ryckelynck, N., and Whaling, P. (2006) 598

Microbial fuel cell energy from an ocean cold seep. Geobiology 4: 123–136.

599

24 Reysenbach, A.-L., Longnecker, K., and Kirshtein, J. (2000) Novel bacterial and archaeal lineages 600

from an in situ growth chamber deployed at a Mid-Atlantic Ridge hydrothermal vent. Appl 601

Environ Microbiol 66: 3798–3806.

602

Salter, S.J., Cox, M.J., Turek, E.M., Calus, S.T., Cookson, W.O., Moffatt, M.F., et al. (2014) Reagent 603

and laboratory contamination can critically impact sequence-based microbiome analyses.

604

BMC Biol 12: 87.

605

Sarradin, P.-M. and Cannat, M. (2014) MOMARSAT2014 cruise,Pourquoi pas ? R/V.

606

Schleucher, J., Schwendinger, M., Sattler, M., Schmidt, P., Schedletzky, O., Glaser, S.J., et al. (1994) 607

A general enhancement scheme in heteronuclear multidimensional NMR employing pulsed 608

field gradients. J Biomol NMR 4: 301–306.

609

Schrenk, M.O., Kelley, D.S., Delaney, J.R., and Baross, J.A. (2003) Incidence and diversity of 610

microorganisms within the walls of an active deep-sea sulfide chimney. Appl Environ 611

Microbiol 69: 3580–3592.

612

Shemesh, M. and Chai, Y. (2013) A Combination of Glycerol and Manganese Promotes Biofilm 613

Formation in Bacillus subtilis via Histidine Kinase KinD Signaling. J Bacteriol 195: 2747–

614

2754.

615

Shen, H.-B., Yong, X.-Y., Chen, Y.-L., Liao, Z.-H., Si, R.-W., Zhou, J., et al. (2014) Enhanced 616

bioelectricity generation by improving pyocyanin production and membrane permeability 617

through sophorolipid addition in Pseudomonas aeruginosa-inoculated microbial fuel cells.

618

Bioresour Technol 167: 490–494.

619

Sokolova, T. and Lebedinsky, A. (2013) CO-Oxidizing Anaerobic Thermophilic Prokaryotes. In 620

Thermophilic Microbes in Environmental and Industrial Biotechnology: Biotechnology of 621

Thermophiles. Satyanarayana, T., Littlechild, J., and Kawarabayasi, Y. (eds). Dordrecht:

622

Springer Netherlands, pp. 203–231.

623

Takai, K., Gamo, T., Tsunogai, U., Nakayama, N., Hirayama, H., Nealson, K.H., and Horikoshi, K.

624

(2004) Geochemical and microbiological evidence for a hydrogen-based, hyperthermophilic 625

subsurface lithoautotrophic microbial ecosystem (HyperSLiME) beneath an active deep-sea 626

hydrothermal field. Extremophiles 8: 269–282.

627

Vorholt, J.A., Hafenbradl, D., Stetter, K.O., and Thauer, R.K. (1997) Pathways of autotrophic CO 2 628

fixation and of dissimilatory nitrate reduction to N 2 O in Ferroglobus placidus. Arch 629

Microbiol 167: 19–23.

630

Vos, P.D. (2015) Bacillales. In Bergey’s Manual of Systematics of Archaea and Bacteria. American 631

Cancer Society, pp. 1–1.

632

Wirth, R. (2017) Colonization of black smokers by hyperthermophilic microorganisms. Trends 633

Microbiol 25: 92–99.

634

Wirth, R., Luckner, M., and Wanner, G. (2018) Validation of a hypothesis: Colonization of black 635

smokers by hyperthermophilic microorganisms. Front Microbiol 9: 524.

636

Yamamoto, M., Nakamura, R., Kasaya, T., Kumagai, H., Suzuki, K., and Takai, K. (2017) 637

Spontaneous and Widespread Electricity Generation in Natural Deep-Sea Hydrothermal 638

Fields. Angew Chem Int Ed 56: 5725–5728.

639

Yilmazel, Y.D., Zhu, X., Kim, K.-Y., Holmes, D.E., and Logan, B.E. (2018) Electrical current 640

generation in microbial electrolysis cells by hyperthermophilic archaea Ferroglobus 641

placidus and Geoglobus ahangari. Bioelectrochemistry 119: 142–149.

642

Zhang, L., Kang, M., Xu, Jiajun, Xu, Jian, Shuai, Y., Zhou, X., et al. (2016) Bacterial and archaeal 643

communities in the deep-sea sediments of inactive hydrothermal vents in the Southwest India 644

Ridge. Sci Rep 6:.

645 646

647

25

Figures

648

649

650

651

Figure 1. Current consumption (red continuous line); pyruvate (blue triangle), glycerol (yellow 652

square) and acetate (green cross) productions over time of culture for each electron-acceptor 653

experiment. The current was obtained from a poised electrode at -590 mV vs SHE for nitrate and 654

sulfate experiments and -300 mV vs SHE for oxygen.

655

656

26 657

Figure 2. Coulombic efficiency for organic products in presence of the different electron 658

acceptors.

659

660

0 10 20 30 40 50 60 70 80 90 100

Nitrate Oxygen Sulfate

Coulombic efficiency (%)

Pyruvate Glycerol Acetate

27 661

Figure 3. Dominant taxonomic affiliation at order level and biodiversity indices of microbial 662

communities from a crushed chimney sample from Capelinhos vent site (Lucky Strike 663

hydrothermal vent field), as plotted on the cathode and liquid media (LM) after the weeks of 664

culture. OTUs representing less than 1% of total sequences of the samples are pooled as ‘Rare 665

OTUs’.

666

667

668

28 Figure 4. Maximum Likelihood phylogenetic tree of archaeal OTUs retrieved on various

669

enrichments on the 293pb 16S fragment obtained in the barcoding 16S method (LM: Liquid 670

Media; WE: Working Electrode, cathode). Numbers at nodes represent bootstrap values inferred 671

by MEGAX. Scale bars represent the average number of substitutions per site.

672

673

674 675

29 676

Figure 5: Schematic representation of microbial colonization of iron-rich hydrothermal

chimney (Capelinhos site on the Lucky Strike hydrothermal field) by

electrolithoautotrophic microorganisms. The production of an abiotic electrical current by

potential differences between the reduced hydrothermal fluid (H

S, metals, CO, CH

, H

…)

and oxidized seawater (O

, SO

, NO

) (Yamamoto et al. 2017) leads to the formation of

electron flux moving towards the chimney surface. This electrons flux can serve directly as

an energy source to enable the growth of electrolithoautotrophic and hyperthermophilic

Archaeolgobales using the CO2

as carbon source and nitrate and/or sulfate as electron

acceptors. In the absence of a usable electron acceptor, Archaeoglobales would be likely

to perform direct interspecies electron transfer to ensure their growth. The electron

acceptor fluctuations, correlated to the continual influx of electric current would favor the

production of organic matters (amino acid, formate, pyruvate, glycerol...) by the

Archaeoglobales. This organic matter is then used by heterotrophic microorganisms by 689

fermentation or respiration (anaerobic or aerobic) thus providing the primal food web

initially present into the hydrothermal ecosystems. The electrical current also could favor

the electrolysis water leading to the abiotic H2 production (not measurable in our abiotic 692

conditions), which would serve as chemical energy source. Arch: Achaeoglobales; Thmc:

693

Thermococcales; Dsfc: Desulfurococcales; Thml: Thermales; Prot: Proteobacteria; Firm:

694

Firmicute; NO3-

: nitrate; SO

: sulfate; O

: dioxygen; CH

: Methane; CO

: Carbon Dioxide;

695

CO: Carbon monoxide; H

S : Hydrogen sulfide ; S°: sulfur; Metals: Fe, Mn, Cu, Zn…

696 697

30 A)

SUPPLEMENTARY INFORMATION 698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

Supplementary Information Figure S1: A) CyclicVoltammograms (scan rate = 20 mV/s) of the abiotic

control, and of the experiments at inoculation time and after 30 days for each condition (Nitrate, Oxygen and Sulfate). B) Reduction peaks extracted from CyclicVoltammograms (scan rate = 20 mV/s) where the baseline have been substracted with the sofware QSoas. The ΔI of reduction peaks are expressed in inversed values.

Cyclovoltammograms carried out with a 3 M Ag/AgCl reference electrode (E= +0.165 V vs SHE at 80°C).

B)

31 718

Supplementary Information Figure S2. Quantification of 16S rRNA gene copies from Bacteria 719

(blue) or Archaea (orange) per gram of crushed chimney, per milliliter of liquid or per cm² of 720

working electrode. Error bar represent the standard deviation obtained on triplicates.

721

32 722

Supplementary Information Figure S3. Heatmap representation of the distribution of dominant 723

OTUs (>0.05%) over the different electron acceptors (LM: Liquid Media; WE: Working Electrode, 724

cathode). OTUs and samples clustering were performed with centroid average method and with 725

Pearson distance measurement method. The red taxa represent the Archaea members and blue 726

taxa, the Bacteria.

727

33 728

729 730 731